Operands could not be broadcast together
I tried executing the command in README and got the following error:
Traceback (most recent call last):
File "demo.py", line 128, in
I also have this problem. It looks like the MFCC preprocess.py is not using the right function. In the paper the authors talked about "We then extract the 13-dimensional MFCC feature, and concatenate it with its first temporal derivatives and log mean energy of volume into the final 28-dimensional audio feature." That explains why the mean_aud and std_aud has the (28, )shape. but the aud we get from the preprocess is default MFCC 13 dimension + two other values as defined in sig2s2mfc_energy() function from preprocess.py
I can't find any code that calculate the "first temporal derivatives". Can somebody show where is the code? thanks.
Same error here, it's a nightmare to reproduce this code
I also have this problem. It looks like the MFCC preprocess.py is not using the right function. In the paper the authors talked about "We then extract the 13-dimensional MFCC feature, and concatenate it with its first temporal derivatives and log mean energy of volume into the final 28-dimensional audio feature." That explains why the mean_aud and std_aud has the (28, )shape. but the aud we get from the preprocess is default MFCC 13 dimension + two other values as defined in sig2s2mfc_energy() function from preprocess.py I can't find any code that calculate the "first temporal derivatives". Can somebody show where is the code? thanks.
I check the source code, but have the same isuue
we must use the special music*.wav as input?
Same error here, it's a nightmare to reproduce this code
@gagolucasm have you dealt with it?
I put a zipped file of the complete demo code on my personal webpage.